The FCC Revised the Satellite Links Starlink and Kuiper Must Protect. The Test That Uses Them Isn't Live Yet.

Astranis sent the FCC a set of reference links for satellites it plans to run at 121 degrees West. The Space Bureau left them out. “We do not include links for prospective operations in the database,” a footnote says [2].

That footnote is in DA 26-973, the notice that revises the list of geostationary (GSO) satellite links every low-orbit constellation has to protect when it shares Ku- and Ka-band downlink spectrum over the U.S. It was released September 11, published in the Federal Register on Thursday, and takes effect December 7 [1, 2]. If you build on satellite broadband, this dataset is where the capacity fight over the next few years gets settled. It is also, for now, a dataset for a test nobody can file.

What the Space Bureau actually changed

Some background. On April 30 the FCC voted to replace the equivalent power-flux density (EPFD) limits, written in the late 1990s, as the way non-geostationary (NGSO) systems protect GSO networks in the 10.7-12.7, 17.3-18.6 and 19.7-20.2 GHz bands [3]. All three commissioners voted for it, according to Via Satellite [6]. The new rule lets an NGSO applicant either certify that it has coordinated with every operating co-frequency GSO network or file a compatibility showing [3]. That showing has to meet four limits against “any GSO reference link”: no more than 3 percent time-weighted average degraded throughput for links using adaptive coding and modulation (ACM), no more than a 0.1 percent absolute change in availability, no more than -10.5 dB I/N for 80 percent of the time for non-ACM links, and a 3-degree avoidance angle of the GSO arc [3].

The Order started with 328 links and told the Space Bureau to revisit them [3]. Nine comments, seven replies and several ex partes later, DA 26-973 does four things [1, 2]:

• Adds links. New ones from DIRECTV, SES, Eutelsat, Hispasat, Astranis and Viasat, plus a replacement set Amazon pulled from FCC licensing records for every authorized Ka-band GSO operator [2].
• Removes links. Out go links for operations outside the U.S., for networks with no current operator, and for networks that never launched or stopped operating over a decade ago. Many came from a 1999 ITU circular letter that predates commercial Ka-band satellites [2].
• Caps a parameter. Some non-ACM links in the initial set had carrier-to-noise (C/N) levels as high as 18 dB [2]. SpaceX proposed a 9.5 dB maximum unavailability threshold, and DIRECTV, which the notice calls “one of the most widespread users of non-ACM links,” supported it for its user links [2]. SES argued that some of its media distribution links run at 12.6 dB to maximize capacity [2]. The Space Bureau sided with SpaceX and DIRECTV, calling the higher values “relative outliers on older networks” [2].
• Trims columns. The dataset drops fields not needed for the showings [2].

It declined to change the selection criteria, including the 25-degree minimum earth station elevation angle that SES and Hispasat opposed [2].

Why one row in a spreadsheet matters

Here’s the mechanism that makes the dataset more important than it looks. The FCC says one protection criterion at one reference link will usually be “the constraining factor” for a given NGSO system, which it calls the “most sensitive link” [2]. Whatever parameters protect that link become the parameters the system uses “when providing service anywhere in the United States” [2]. The three levers are the number of co-frequency beams serving a location, the GSO-arc avoidance angle, and power [2]. The reason it’s national is practical: NGSO operators generally don’t know where blanket-licensed GSO earth stations sit [2].

So the dataset is worst-case engineering. A single bad row, say a legacy link with an 18 dB threshold that nobody operates, could set the beam count for every Starlink or Kuiper terminal in the country. That’s why the cleanup is the real news here, and why the GSO side cared so much about it. SES argued that legacy links deserved deference; the Space Bureau answered that no current operator had claimed them [2]. It also said it is not setting up “a mechanism for future removals” [2]. Additions, though, stay open, and the Order says operating NGSO systems “may be required to adjust their operations” to protect new links [3].

Put the Astranis footnote next to that sentence. A GSO newcomer can’t get its planned links into the protected set. Once it’s operating, its links can come in, and if one of them becomes the most sensitive link, every NGSO operator’s limits move. I think that’s a reasonable design, and I’d rather have it than a list frozen in 1999. But it means the capacity your satellite backhaul or connected-vehicle plan assumes isn’t a fixed number. It floats with a spreadsheet that the Space Bureau can revise after comment [3].

The GSO operators’ broader case shouldn’t get lost. SES told the FCC the new framework would shift interference burdens onto GSO operators and could devalue GSO assets to the point of stranding them; the FCC said SES gave no economic evidence for that [3]. Via Satellite reported that SES and Eutelsat asked for a further proceeding on an aggregate NGSO interference limit, and that DIRECTV sought a larger avoidance angle [6]. DIRECTV has a petition for reconsideration of the Order pending; oppositions were due July 21 [5]. And the Space Bureau itself concedes the framework “has yet to be implemented” and that it will reassess protection criteria with “real-world operational information” [2].

The part that isn’t live

The Order took effect July 13 except for 47 CFR 25.146(a)(3) and 25.289(a)(2), which are “indefinitely delayed” [3]. Section 25.146(a)(3) is the compatibility showing. The Space Bureau says that, until it gets Paperwork Reduction Act approval, it “has not yet begun reviewing or approving compatibility showings under the new rules” [2]. The FCC’s PRA notice covering those showings, under OMB Control No. 3060-0678, asks for comments by November 9 [4].

So on December 7 the reference links become effective for a procedure that can’t be used yet. That’s not a scandal. Paperwork approval is ordinary process, and federal law bars penalties for failing to comply with a collection that lacks a valid OMB control number [8]. It is a gap, though, between an Order the FCC and OMB’s regulatory office classified as a major rule (one likely to have an annual effect on the economy of $100 million or more) and a working filing path [3, 7]. The FCC put the benefits at “well over $2 billion,” with capacity increases of 100 to 700 percent on the same satellites [3]. None of that arrives through this route until OMB signs off.

Back in October I wrote about the upper C-band transition plans, where three satellite operators’ filings, SES’s among them, set the bill for bidders. Same pattern here: the policy is decided, and the numbers inside the filings decide what it’s worth.

What I’m watching

• November 9 and after. PRA comments on OMB 3060-0678 close [4]. Next comes OMB review and a Federal Register notice setting an effective date for 25.146(a)(3) [3]. Until then, no showings.
• DIRECTV’s reconsideration petition. Whether the FCC rules on it before the first showings are filed [5].
• The first new link. Whether a GSO operator, Astranis or anyone else, asks to add a link once it’s operating, and whether that link becomes the most sensitive one [2, 3].